[ds-fusion] Get While loop analysis with copy fusion #21375

shraiysh · 2025-01-13T18:17:03Z

In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index.

Imported from GitHub PR #21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- ae85690 by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Merging this change closes #21375 FUTURE_COPYBARA_INTEGRATE_REVIEW=#21375 from shraiysh:while_loop_analysis ae85690 PiperOrigin-RevId: 715047991

Imported from GitHub PR openxla/xla#21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- ae85690876a106c4d74715fed299779e29e8e641 by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Merging this change closes #21375 Reverts 49b2a87 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#21375 from shraiysh:while_loop_analysis ae85690876a106c4d74715fed299779e29e8e641 PiperOrigin-RevId: 715047991

Imported from GitHub PR #21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- ae85690 by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Merging this change closes #21375 FUTURE_COPYBARA_INTEGRATE_REVIEW=#21375 from shraiysh:while_loop_analysis ae85690 PiperOrigin-RevId: 715047991

Imported from GitHub PR openxla/xla#21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- ae85690876a106c4d74715fed299779e29e8e641 by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Merging this change closes #21375 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#21375 from shraiysh:while_loop_analysis ae85690876a106c4d74715fed299779e29e8e641 PiperOrigin-RevId: 715047991

Imported from GitHub PR openxla/xla#21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- ae85690876a106c4d74715fed299779e29e8e641 by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Merging this change closes #21375 PiperOrigin-RevId: 715080653

fhoushmand · 2025-01-13T23:07:06Z

Rolling back due to internal test breakage.

shraiysh · 2025-01-13T23:22:52Z

Do we have a reduced testcase or the reason for this breakage? I can investigate and try to submit again.

github-actions · 2025-01-14T01:12:35Z

This PR was rolled back in 3aa5d48!

Imported from GitHub PR #21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- 0a7551b by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Merging this change closes #21375 FUTURE_COPYBARA_INTEGRATE_REVIEW=#21375 from shraiysh:while_loop_analysis 0a7551b PiperOrigin-RevId: 715237085

Imported from GitHub PR openxla/xla#21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- 0a7551bbba67f5900eb0bb5fc4d193f37d5b63ee by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Merging this change closes #21375 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#21375 from shraiysh:while_loop_analysis 0a7551bbba67f5900eb0bb5fc4d193f37d5b63ee PiperOrigin-RevId: 715237085

dimitar-asenov · 2025-01-14T12:56:18Z

I'm working on trying to get a minimal reproducer. In the mean time, here's then failing stack trace:

Check failed at xla/hlo/ir/hlo_computation.cc:1413: DCHECK(changed) 

xla/hlo/ir/hlo_computation.cc:1413 xla::HloComputation::ReplaceInstruction()
xla/hlo/ir/hlo_computation.cc:1381 xla::HloComputation::ReplaceWithNewInstruction()
xla/tools/hlo_extractor.cc:332 xla::(anonymous namespace)::Inline()
xla/tools/hlo_extractor.cc:369 xla::ExtractModule()
xla/hlo/analysis/while_loop_analysis.cc:173 xla::GetUniqueGTEDependenceIndex()
xla/hlo/analysis/while_loop_analysis.cc:432 xla::GetLoopInductionVarTupleIdx()
xla/hlo/analysis/while_loop_analysis.cc:487 xla::MatchTrivialLoopRange()

Enabling extra logging reveals that CHECK fails happens because the replacement fails with:

hlo_computation.cc:1432 Skipping replacement because old instruction has control dependencies

In this particular case the old instruction is a fusion and the new one is a call.

Will share a reproducer once I get a more minimal example.

FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#21375 from shraiysh:while_loop_analysis a435fbd2eadc17269d7bccbe141dcf7a21cc20e8 PiperOrigin-RevId: 720702570

FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#21375 from shraiysh:while_loop_analysis a435fbd2eadc17269d7bccbe141dcf7a21cc20e8 PiperOrigin-RevId: 720696248

…ks using XLA's FFI. FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#21375 from shraiysh:while_loop_analysis a435fbd2eadc17269d7bccbe141dcf7a21cc20e8 PiperOrigin-RevId: 720649341

…ions. This is in preparation for a larger change, so that `_check_arrays` can be called before Array creation in XLA and the user gets more helpful JAX error messages instead of XLA errors. FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#21375 from shraiysh:while_loop_analysis a435fbd2eadc17269d7bccbe141dcf7a21cc20e8 PiperOrigin-RevId: 720692284

FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#21375 from shraiysh:while_loop_analysis a435fbd2eadc17269d7bccbe141dcf7a21cc20e8 PiperOrigin-RevId: 718029356

FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#21375 from shraiysh:while_loop_analysis a435fbd2eadc17269d7bccbe141dcf7a21cc20e8 PiperOrigin-RevId: 716201082

Imported from GitHub PR openxla/xla#21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- 3147ec926aa1c6fdfa2f4376668434c9a2fbeb87 by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. -- a435fbd2eadc17269d7bccbe141dcf7a21cc20e8 by Shraiysh Vaishay <[email protected]>: Relay control dependencies while converting fusion to call (extractor) Merging this change closes #21375 PiperOrigin-RevId: 720691268

github-actions · 2025-01-29T08:38:04Z

This PR was rolled back in 95c45c6!

akuegel · 2025-01-29T09:04:22Z

@shraiysh FYI this triggered a bug in AlgebraicSimplifier, by now I have a fix ready, but we reverted first to unbreak the user as soon as possible. After my fix has landed we will attempt to reland.

dimitar-asenov · 2025-01-29T15:29:26Z

The fix has landed and this PR was re-merged. Let's hope there are no more breakages.

shraiysh · 2025-01-29T17:28:14Z

Thanks for helping land this @akuegel @dimitar-asenov! Let me know if I can help if there are future breakages.

dimitar-asenov · 2025-01-31T20:52:29Z

Unfortunately, we'll have to roll this back again as it caused another preexisting issue with the Algebraic Simplifier to surface. We'll look into it next week, I don't have a reproducer at the moment.

shraiysh · 2025-01-31T22:51:24Z

Unfortunately, we'll have to roll this back again as it caused another preexisting issue with the Algebraic Simplifier to surface. We'll look into it next week, I don't have a reproducer at the moment.

Sure. No problem. Please let me know how I can help.

Imported from GitHub PR #21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- 3147ec9 by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. -- a435fbd by Shraiysh Vaishay <[email protected]>: Relay control dependencies while converting fusion to call (extractor) Merging this change closes #21375 FUTURE_COPYBARA_INTEGRATE_REVIEW=#21375 from shraiysh:while_loop_analysis a435fbd PiperOrigin-RevId: 722709283

Imported from GitHub PR openxla/xla#21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- 3147ec926aa1c6fdfa2f4376668434c9a2fbeb87 by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. -- a435fbd2eadc17269d7bccbe141dcf7a21cc20e8 by Shraiysh Vaishay <[email protected]>: Relay control dependencies while converting fusion to call (extractor) Merging this change closes #21375 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#21375 from shraiysh:while_loop_analysis a435fbd2eadc17269d7bccbe141dcf7a21cc20e8 PiperOrigin-RevId: 722709283

Imported from GitHub PR #21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- 3147ec9 by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. -- a435fbd by Shraiysh Vaishay <[email protected]>: Relay control dependencies while converting fusion to call (extractor) Merging this change closes #21375 FUTURE_COPYBARA_INTEGRATE_REVIEW=#21375 from shraiysh:while_loop_analysis a435fbd PiperOrigin-RevId: 722709283

Imported from GitHub PR openxla/xla#21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- 3147ec926aa1c6fdfa2f4376668434c9a2fbeb87 by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. -- a435fbd2eadc17269d7bccbe141dcf7a21cc20e8 by Shraiysh Vaishay <[email protected]>: Relay control dependencies while converting fusion to call (extractor) Merging this change closes #21375 Reverts changelist 723246423 FUTURE_COPYBARA_INTEGRATE_REVIEW=openxla/xla#21375 from shraiysh:while_loop_analysis a435fbd2eadc17269d7bccbe141dcf7a21cc20e8 PiperOrigin-RevId: 722709283

dimitar-asenov · 2025-02-05T10:29:56Z

This change was re-merged. Let's hope that was the last issue.

723477680 by A. Unique TensorFlower<[email protected]>: [XLA] Tag timeout tests as `not_run:arm` Similarly to cl/722883015 tagging also: * //third_party/tensorflow/compiler/xla/python/transfer:socket_bulk_transport_test * //third_party/tensorflow/compiler/xla/python/transfer:socket-server_test * //third_party/tensorflow/compiler/xla/python/transfer:event_loop_test -- 723476384 by A. Unique TensorFlower<[email protected]>: Parse XLA_FLAGS environment variable every time, conditionally on xla_flags_reset flag. -- 723471749 by A. Unique TensorFlower<[email protected]>: [XLA:GPU] Rename `IsSyncCollective` and move to a GPU specific file. The implementation is specific to the GPU backend. -- 723470593 by A. Unique TensorFlower<[email protected]>: [XLA:GPU] move DotDecompose out of simplification pipeline That seems to be a better approach then moving TransposeFold to simplification-2 in 961e5c25fbd4082a1ac4f2e0865ad28163d12f7d: 1. There is a report that previous change has resulted in perf degradation openxla/xla#22081 2. I have found another case when DotDecompose is competing with algsimp. Added a test for that. Overall, having an pass that expands operation together with passes that are trying to do the simplification asks for such infinite loops. --- For archeologists: passes DotDimensionSorter and DotDecomposer were added along with GpuAlgebraicSimplifier as it previously could have added multiple contracting dimensions to dot. But cudnn does not support dots with 2+ dimensions, forcing us to use a less efficient loop emitter. - That what "// AlgebraicSimplifier may add contracting dimensions to a dot." comment was about. After a while simplifier started to use supports_non_canonical_dots to guard against this case. So it should be safe to remove dot decomposer and friends. -- 723469960 by A. Unique TensorFlower<[email protected]>: PR #22334: [ROCm] Fix flaky gpu compiler test when building with rocm Imported from GitHub PR openxla/xla#22334 This change fixes the flaky gpu compiler test used to run on rocm CI pipeline gate. Triton pipeline was wrongly using the TritonGPUAccelerateMatmul pass which supports cuda only. In rocm there is a different pass which is now used in the rocm pipeline. https://github.com/triton-lang/triton/blob/main/third_party/amd/lib/TritonAMDGPUTransforms/AccelerateAMDMatmul.cpp Copybara import of the project: -- c5f600f03aa87d155bb624bedb0584e635af190e by Alexandros Theodoridis <[email protected]>: Fix flaky gpu compiler test when building with rocm Merging this change closes #22334 -- 723453199 by A. Unique TensorFlower<[email protected]>: Automated Code Change -- 723445422 by A. Unique TensorFlower<[email protected]>: Automated Code Change -- 723443292 by A. Unique TensorFlower<[email protected]>: [pjrt] Removed deprecated `PjRtBuffer::CopyToDevice` -- 723434255 by A. Unique TensorFlower<[email protected]>: Automated Code Change -- 723430683 by A. Unique TensorFlower<[email protected]>: Automated Code Change -- 723426786 by A. Unique TensorFlower<[email protected]>: PR #22258: [GPU][NFC] Avoid always printing complete PGLE profiles. Imported from GitHub PR openxla/xla#22258 Copybara import of the project: -- 025352635a155e447559d83c471369559aad5981 by Ilia Sergachev <[email protected]>: [GPU][NFC] Avoid always printing complete PGLE profiles. Merging this change closes #22258 -- 723426773 by A. Unique TensorFlower<[email protected]>: PR #21375: [ds-fusion] Get While loop analysis with copy fusion Imported from GitHub PR openxla/xla#21375 In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. Copybara import of the project: -- 3147ec926aa1c6fdfa2f4376668434c9a2fbeb87 by Shraiysh Vaishay <[email protected]>: [ds-fusion] Get While loop analysis with copy fusion In later stages of optimization, there are instances of copy fusion on the parameter of the while body. With this, we need to allow inlining of fusions while getting the induction variable index, otherwise we cannot deduce the tuple index. -- a435fbd2eadc17269d7bccbe141dcf7a21cc20e8 by Shraiysh Vaishay <[email protected]>: Relay control dependencies while converting fusion to call (extractor) Merging this change closes #21375 -- 723425710 by A. Unique TensorFlower<[email protected]>: [XLA] Add const reference versions of `ForEachInstructionWithPred` and `ForEachInstructionWithOpcode`. These are more permissive and semantically equivalent. -- 723425622 by A. Unique TensorFlower<[email protected]>: Remove dead code (NFC) We compute the total number of tiles in a variable `num_tiles` but then never use it. So remove it. -- 723419822 by A. Unique TensorFlower<[email protected]>: Automated Code Change -- 723402058 by A. Unique TensorFlower<[email protected]>: compat: Update forward compatibility horizon to 2025-02-05 -- 723401869 by A. Unique TensorFlower<[email protected]>: Update GraphDef version to 2129. -- 723396271 by A. Unique TensorFlower<[email protected]>: [XLA] Support different operand and result types in AlgebraicSimplifierVisitor::HandlePad. I checked that none of the other cases in HandlePad require any adjustments. -- 723389764 by A. Unique TensorFlower<[email protected]>: Automated Code Change -- 723370461 by A. Unique TensorFlower<[email protected]>: Use matchers_oss in vendor code -- 723367856 by A. Unique TensorFlower<[email protected]>: Update users of TSL headers and targets to new location in XLA Updating: - `env.h` - `env_time.h` - `errors.h` - `file_statistics.h` - `file_system.h` - `file_system_helper.h` - `logging.h` - `macros.h` - `status.h` - `status_matchers.h` - `status_to_from_proto.h` - `statusor.h` - `test.h` - `test_benchmark.h` - `threadpool.h` - `threadpool_async_executor.h` - `threadpool_interface.h` - `threadpool_options.h` - `types.h` and associated targets. -- 723349025 by A. Unique TensorFlower<[email protected]>: Fix inference request analysis aggregated on batch size, by aggregating only the requests included in a single batch, as large request split into multiple batches will introduce confusing results (eg. the device time will be the sum of the 2 batch processing). -- 723344172 by A. Unique TensorFlower<[email protected]>: Automated Code Change -- 723340771 by A. Unique TensorFlower<[email protected]>: Automated Code Change -- 723337100 by A. Unique TensorFlower<[email protected]>: Automated Code Change -- 723321370 by A. Unique TensorFlower<[email protected]>: Stop modifying the TraceEventsContainer in DoStoreAsLevelDbTable. This behavior is not intuitive (modifying a const value that was passed in) and unnecessary. -- 723307829 by A. Unique TensorFlower<[email protected]>: Automated rollback of changelist 723246423. 723278167 by A. Unique TensorFlower<[email protected]>: Update users of TSL headers and targets to new location in XLA Updating: - `env.h` - `env_time.h` - `errors.h` - `file_statistics.h` - `file_system.h` - `file_system_helper.h` - `logging.h` - `macros.h` - `status.h` - `status_matchers.h` - `status_to_from_proto.h` - `statusor.h` - `test.h` - `test_benchmark.h` - `threadpool.h` - `threadpool_async_executor.h` - `threadpool_interface.h` - `threadpool_options.h` - `types.h` and associated targets. -- 723265881 by A. Unique TensorFlower<[email protected]>: Add the list of Qualcomm SoCs supporting NPU. -- 723248792 by A. Unique TensorFlower<[email protected]>: Add Q/DQ annotation lowering support. LowerQuantAnnotationsPass now supports quant.quantize and quant.dequantize composite lowering. These patterns make adjustments to the function signatures if necessary. -- 723246423 by A. Unique TensorFlower<[email protected]>: PR #85476: Support Qnn Wrappers for LiteRt Imported from GitHub PR #85476 # WHAT - Basic wrapper for QNN types, handle dynamic resources along with wrapper instances. - Make these wrappers independent to LiteRT/tflite - Only depend on QNN and STL ### `ScalarParamWrapper` - Wrap `Qnn_Param_t` with `QNN_PARAMTYPE_SCALAR` for `paramType` - Choose correct `QNN_DATATYPE` based on the data type ### `TensorParamWrapper` - Wrap `Qnn_Param_t` with `QNN_PARAMTYPE_TENSOR` for `paramType` ### `UndefinedQuantizeParamsWrapper` - Wrap `Qnn_QuantizeParams_t` - Default for quantization parameter ### `ScaleOffsetQuantizeParamsWrapper` - Wrap `Qnn_QuantizeParams_t` for per-tensor quantization ### `AxisScaleOffsetQuantizeParamsWrapper` - Wrap `Qnn_QuantizeParams_t` for per-axis quantization ### `TensorWrapper` - Wrap `Qnn_TensorType_t` - Handle dynamic resource, e.g. name, dimensions, weight data. ### `OpWrapper` - Wrap `Qnn_OpConfig_t` - Handle dynamic resource, e.g. name, input output tensors, params Copybara import of the project: -- 4833a20 by weilhuan-quic <[email protected]>: LiteRt Qualcomm wrappers -- 725f571 by weilhuan-quic <[email protected]>: TensorWrapper GetDataTypeSize() return bytes instead of bits -- dd3f251 by weilhuan-quic <[email protected]>: comment qnn_lib_headers -- 06e0616 by weilhuan-quic <[email protected]>: Change license Merging this change closes #85476 -- PiperOrigin-RevId: 723477680

shraiysh requested a review from bchetioui January 13, 2025 18:17

shraiysh self-assigned this Jan 13, 2025

shraiysh added the kokoro:force-run Forces CI to rerun label Jan 13, 2025

kokoro-team removed the kokoro:force-run Forces CI to rerun label Jan 13, 2025

dimitar-asenov requested a review from fhoushmand January 13, 2025 19:24

fhoushmand approved these changes Jan 13, 2025

View reviewed changes

copybara-service bot mentioned this pull request Jan 13, 2025

PR #21375: [ds-fusion] Get While loop analysis with copy fusion #21376

Merged

copybara-service bot mentioned this pull request Jan 13, 2025

PR #21375: [ds-fusion] Get While loop analysis with copy fusion tensorflow/tensorflow#84786

Merged

copybara-service bot closed this in 07223ca Jan 13, 2025

shraiysh mentioned this pull request Jan 13, 2025

[ds-fusion] Add runtime support for host calculation of offsets in ds fusion #20332

Closed

shraiysh reopened this Jan 13, 2025

shraiysh force-pushed the while_loop_analysis branch from ae85690 to 0a7551b Compare January 14, 2025 06:16

shraiysh requested a review from fhoushmand January 14, 2025 06:16

shraiysh added the kokoro:force-run Forces CI to rerun label Jan 14, 2025

kokoro-team removed the kokoro:force-run Forces CI to rerun label Jan 14, 2025

copybara-service bot mentioned this pull request Jan 14, 2025

PR #21375: [ds-fusion] Get While loop analysis with copy fusion #21395

Closed

copybara-service bot mentioned this pull request Jan 14, 2025

PR #21375: [ds-fusion] Get While loop analysis with copy fusion tensorflow/tensorflow#84815

Closed

shraiysh force-pushed the while_loop_analysis branch from 0a7551b to 2d67160 Compare January 14, 2025 18:07

copybara-service bot mentioned this pull request Jan 28, 2025

Actually copy the extra wheels into the expected directory afterwards. tensorflow/tensorflow#85982

Merged

copybara-service bot mentioned this pull request Jan 28, 2025

Implement Jax CPU/GPU callbacks with XLA's FFI. tensorflow/tensorflow#85983

Merged

copybara-service bot mentioned this pull request Jan 28, 2025

Refactor Jax FFI lowering to prepare for implementing CPU/GPU callbacks using XLA's FFI. tensorflow/tensorflow#85984

Closed

copybara-service bot mentioned this pull request Jan 28, 2025

Return arrays from ArrayImpl._check_and_rearrange. tensorflow/tensorflow#85985

Merged

copybara-service bot mentioned this pull request Jan 28, 2025

Improve MSA readabilty by changing some auto types to their type name. tensorflow/tensorflow#85422

Merged

copybara-service bot mentioned this pull request Jan 28, 2025

[xla:pjrt] Pass host callback pointers to FFI handlers via FFI's ExecutionContext. tensorflow/tensorflow#85987

Merged

shraiysh reopened this Feb 3, 2025

copybara-service bot mentioned this pull request Feb 3, 2025

PR #21375: [ds-fusion] Get While loop analysis with copy fusion #22244

Merged

copybara-service bot mentioned this pull request Feb 3, 2025

PR #21375: [ds-fusion] Get While loop analysis with copy fusion tensorflow/tensorflow#86470

Closed

copybara-service bot closed this in 581538c Feb 5, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[ds-fusion] Get While loop analysis with copy fusion #21375

[ds-fusion] Get While loop analysis with copy fusion #21375

Uh oh!

shraiysh commented Jan 13, 2025

Uh oh!

fhoushmand commented Jan 13, 2025

Uh oh!

shraiysh commented Jan 13, 2025

Uh oh!

github-actions bot commented Jan 14, 2025

Uh oh!

dimitar-asenov commented Jan 14, 2025

Uh oh!

github-actions bot commented Jan 29, 2025

Uh oh!

akuegel commented Jan 29, 2025

Uh oh!

dimitar-asenov commented Jan 29, 2025

Uh oh!

shraiysh commented Jan 29, 2025

Uh oh!

dimitar-asenov commented Jan 31, 2025

Uh oh!

shraiysh commented Jan 31, 2025

Uh oh!

dimitar-asenov commented Feb 5, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

[ds-fusion] Get While loop analysis with copy fusion #21375

[ds-fusion] Get While loop analysis with copy fusion #21375

Uh oh!

Conversation

shraiysh commented Jan 13, 2025

Uh oh!

fhoushmand commented Jan 13, 2025

Uh oh!

shraiysh commented Jan 13, 2025

Uh oh!

github-actions bot commented Jan 14, 2025

Uh oh!

dimitar-asenov commented Jan 14, 2025

Uh oh!

github-actions bot commented Jan 29, 2025

Uh oh!

akuegel commented Jan 29, 2025

Uh oh!

dimitar-asenov commented Jan 29, 2025

Uh oh!

shraiysh commented Jan 29, 2025

Uh oh!

dimitar-asenov commented Jan 31, 2025

Uh oh!

shraiysh commented Jan 31, 2025

Uh oh!

dimitar-asenov commented Feb 5, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants